These were allowed to stand on the incubator at 33°–35°C., and examined for acid each day by the method described in the previous communication for starch testing. They all remained neutral, although the bacteria developed in all but No. 1. These experiments were repeated several times, with the same result in every case. They show that this ferment is unable to act on starch either in its insoluble or soluble condition, alone or in the presence of nitrogenous bodies. Now it has been known for a considerable period that bran contains an unorganised ferment called _cerealin_,[171] which is capable of changing starch into dextrin and other carbohydrates; but the information to be obtained about it was very meagre, and as it appeared that this body might play an important part in the fermentation, we proceeded to prepare some pure cerealin and to ascertain its action on pure starch. The cerealin was prepared by taking a kilo of bran and extracting it with 2 litres of distilled water at 30° C; the extract was filtered clear and 2 litres of strong spirits of wine containing 90 per cent. alcohol was added, when a flocculent precipitate separated, which was washed on a filter with alcohol, dehydrated with absolute alcohol, and dried over H_{2}SO_{4}. The cerealin thus prepared is an amorphous substance not quite white, difficultly soluble in water, though we think this is due to its having been coagulated, and that it might be prepared in some other way which would show it to be more soluble than that which you now see. [171] Watts’ Dict., Old Ed. Cerealin discovered by Mèje Mouriés, Comptes rend. 37, 351; 38, 505; 43, 1122; 48, 431; 50, 467. To show its action on starch we took 10 grm. of pure starch in 200 c.c. water at 40° C., and placed equal quantities in two flasks; to No. 1 about 0·1 grm. of the cerealin was added; No. 2 was left blank. These were kept at 40° C. for 10 hours, the clear liquid filtered off, and examined with Fehling’s solution. No. 1 reduced it strongly, showing that _glucoses_ were present in considerable quantity. No. 2 had no effect whatever. Addition of alcohol to No. 1 gave a white precipitate; this was thrown on to a filter, washed with alcohol, and dried at 100°, again dissolved in water, and inverted by boiling the solution with 1/20 part of strong sulphuric acid and then neutralising with sodium hydrate. The resulting solution reduced Fehling’s solution, showing that the body was dextrin. Strong bran infusion (which of course contains cerealin) acting on a thick starch mucilage, liquefies it, and forms glucoses and dextrin. We have thus shown that the cerealin produces _glucoses_ as well as dextrin, both from solid starch and from starch mucilage. This is most important, as it has been previously shown by one of us that the ferment attacks glucose very easily.